Cancer Causes Control (2014) 25:321–328 DOI 10.1007/s10552-013-0333-7
ORIGINAL PAPER
Body weight in early adulthood, adult weight gain, and risk of endometrial cancer in women not using postmenopausal hormones Victoria L. Stevens • Eric J. Jacobs • Alpa V. Patel • Juzhong Sun • Susan M. Gapstur Marjorie L. McCullough
•
Received: 18 September 2013 / Accepted: 20 December 2013 / Published online: 1 January 2014 Ó Springer Science+Business Media Dordrecht 2013
Abstract Purpose High body mass index (BMI) measured in middle age or later is an established risk factor for endometrial cancer. However, whether BMI measured in early adulthood and adult weight change are associated with endometrial cancer risk is less clear, particularly among nonusers of postmenopausal hormones (PMH). Methods These associations were investigated among women in the Cancer Prevention Study II Nutrition Cohort. Women taking PMH (n = 11,624, 12 % of all women) were excluded, and the analysis was limited to 33,057 postmenopausal women who did not take PMH. Between enrollment in 1992/1993 and 30 June 2009, 447 women were diagnosed with endometrial cancer. Cox proportional hazards regression was used to calculate hazard rate ratios (RR) and corresponding 95 % confidence intervals (CI) for the association of BMI at age 18, calculated from recalled weight, and weight change between age 18 and 1992, with endometrial cancer incidence. Results BMI at age 18 was associated with higher risk of endometrial cancer in multivariable models adjusted for other risk factors (RR 1.29, 95 % CI 1.12–1.49 per 5 BMI units). Similarly, adult weight change was associated with higher risk of endometrial cancer (RR 1.81, 95 % CI 1.66–1.98 per 5 BMI unit change) in multivariable models adjusted for other risk factors.
Electronic supplementary material The online version of this article (doi:10.1007/s10552-013-0333-7) contains supplementary material, which is available to authorized users. V. L. Stevens (&) E. J. Jacobs A. V. Patel J. Sun S. M. Gapstur M. L. McCullough Epidemiology Research Program, American Cancer Society, 250 Williams St, NW, Atlanta, GA 30329, USA e-mail: [email protected]
Conclusions High BMI at age 18 and greater adult weight gain were strongly associated with risk of endometrial cancer. These results underscore the importance of both avoiding overweight/obesity in young adulthood and preventing weight gain thereafter to minimize risk of this cancer. Keywords Endometrial cancer Obesity Weight change Postmenopausal
Introduction Endometrial cancer is the most common gynecologic cancer and fourth most common cancer among women in the USA [1]. Obesity is an established risk factor for this cancer; it is estimated that risk increases about 59 % for every 5 unit increase in body mass index (BMI, kg/m2) [2], and overweight and obesity are responsible for 57 % of all cases of endometrial cancer in the USA [3]. Like most risk factors for this cancer, obesity increases exposure to estrogen unopposed by progesterone in both premenopausal and postmenopausal women [4]. With the dramatic increases in the prevalence of overweight and obesity in both children [5] and adults [6] in recent decades, a clear understanding of the consequences of excess bodyweight early in life and weight gain as an adult is essential knowledge for physicians to appropriately advise their patients as to how to minimize cancer risk. The studies which established obesity as a risk factor for endometrial cancer measured BMI at or near the time of diagnosis [7], which is usually after the age of 50. Despite numerous epidemiologic studies [8–27], the associations of BMI in early adulthood and adult weight change with endometrial cancer are not well defined. Many of the
123
322
previous studies only investigated one of these exposures [8, 9, 11, 18, 20, 22, 25], and several of the studies of both BMI in early adulthood and adult weight change [12, 14– 16, 21, 27] adjusted for later life BMI in their statistical model, which may obscure the contribution of these factors to increased risk of endometrial cancer. Five [13, 17, 19, 23, 24] of the six [10] studies that did not adjust for later life BMI reported statistically significant positive associations of adult weight change with endometrial cancer risk, but the findings for early adulthood BMI were mixed with three studies reporting a positive association [17, 23, 24] and three reporting null results [10, 13, 19]. An additional consideration for understanding the association of BMI at early adulthood and adult weight change with endometrial cancer risk is that the use of postmenopausal hormone therapy (PMH) significantly attenuates the association between obesity and endometrial cancer [20, 21, 28]. With one exception [26], all previous studies of BMI in early adulthood and/or adult weight change with endometrial cancer risk [8–25, 27] included both users and nonusers of PMH. Since publication of results from the Women’s Health Initiative (WHI) [29], the prevalence of PMH had been declining steadily [30]. Because future populations of women will include few long-term PMH users, it is important to examine the relative risks of early adult BMI and adult weight change associated with endometrial cancer in non-PMH users. We examined the associations of BMI at age 18 and adult weight change with endometrial cancer in the American Cancer Society’s Cancer Prevention Study II (CPS-II) Nutrition Cohort, a large prospective study of older adults. These analyses were conducted among women not using PMH so that the results would be applicable to women currently seeking advice on minimizing cancer risk.
Materials and methods Study population Subjects for this study were female participants from the CPS-II Nutrition Cohort, a prospective study of cancer incidence and mortality among 86,404 men and 97,785 women. The Nutrition Cohort, which is described in detail elsewhere [31], was initiated in 1992 as a subcohort of CPS-II, a prospective study of cancer mortality begun in 1982 involving approximately 1.2 million Americans. Nutrition Cohort participants were recruited from CPS-II members who resided in 21 states and were predominantly between the ages of 50 and 74 years. Participants completed a self-administered questionnaire at baseline that included demographic, medical, anthropometric, dietary,
123
Cancer Causes Control (2014) 25:321–328
and lifestyle information. Follow-up questionnaires were sent to all living Nutrition Cohort members in 1997 and every two years thereafter to update exposure information and to ascertain newly diagnosed cancers. The response rate on all of the follow-up questionnaires, among those cohort participants who were mailed surveys, was at least 87 %. Follow-up was from the date of receipt of the baseline survey in 1992/1993 until the date of diagnosis of endometrial cancer, death, date of the last survey returned, or 30 June 2009 whichever came first. All aspects of the CPS-II Nutrition Cohort study are approved by the Emory University Institutional Review Board. Women were excluded from this analysis if they were lost to follow-up (n = 3,111), reported any prevalent cancer other than non-melanoma skin cancer in either 1982 (n = 6,413) or 1992 (n = 6,681), or self-reported endometrial cancer on their first follow-up survey which could not be verified (n = 45). Women who reported endometrial cancer during later follow-up that could not be verified were censored on the date of their last cancer-free survey. Also excluded were women with missing BMI at age 18 or in 1992 and who reported a BMI \ 10 kg/m2 at age 18 or BMI \ 18.5 kg/m2 in 1992 (n = 3,658). Women who were premenopausal or whose menopausal status was unknown (n = 4,124) and who reported that their uterus had been removed or had unknown uterine status at baseline (n = 28,943) were also excluded. Finally, because postmenopausal hormone (PMH) use is known to modify the association between BMI and risk of endometrial cancer, all women who reported current use of PMH or for whom use of PMH was unknown in 1992 (n = 11,753) were also excluded. Additionally, women who started taking PMH after baseline were censored on the date of receipt of the questionnaire on which this use was reported. Women who did not return two consecutive surveys but returned a later survey were censored 4 months after the second missed survey was mailed, and women who underwent hysterectomy after enrollment were censored at the date of receipt of the questionnaire on which this procedure was reported. The final analytic cohort for this study included 33,057 women. Endometrial cancer cases After all exclusions, we identified and verified 447 cases of endometrial cancer diagnosed between enrollment in 1992/1993 and 30 June 2009 for this analysis. These cases (ICD-0 code 54.1) were initially identified through selfreport on the follow-up questionnaires and subsequently verified through medical records (n = 286) or linkage with state cancer registries (n = 161) [32]. Endometrial cancer cases were further subtyped based on histology into endometrioid (type I) cancers (n = 296)
Cancer Causes Control (2014) 25:321–328
and high grade, poor prognosis (type II) cancers (n = 121). Type II cancers included serous papillary, clear cell, or squamous cell histology (n = 34), and any additional International Federation of Gynecology and Obstetrics grade three or four cancers (n = 87) [33]. Cases with missing grade (n = 30) were excluded from subtype analyses. BMI and adult weight change Women self-reported their current weight and weight at age 18 on the Nutrition Cohort baseline questionnaire in 1992. These weights were used to calculate BMI by dividing by the square of their height, which was ascertained in 1982. Weight change between age 18 and 1992 was determined from the difference between two values given for each of these times and is expressed as BMI change. The rate of BMI change from age 18 to 1992 was determined by dividing the BMI change by 10 years. Statistical analyses All statistical analyses were conducted with SAS v9.3. Cox proportional hazards regression [34] was used to calculate hazard rate ratios (RR) and corresponding 95 % confidence intervals (CI) for the association of BMI and weight change, modeled as categorical and continuous variables, with endometrial cancer incidence. p values for linear trend were determined using continuous variables for BMI and weight change. The Cox proportional hazards assumption was tested by modeling a multiplicative interaction term between BMI at 18 or weight change and time. No violations were found. Interactions between BMI at age 18 and weight change were tested using the likelihood ratio test to determine whether the association of weight gain differed between women of different body sizes at early adulthood. All Cox models were adjusted for age by stratifying on exact year of age in 1992. Additional covariates included in the multivariable-adjusted models were family history of endometrial cancer (yes, no), alcohol consumption (not current drinker, \1 drink/week, 1–6 drinks/week, 1 drink/ day, C2 drinks/day, missing), smoking status (never, former \20 years, former C20 years, current \40 years, current C40 years, ever/unclassifiable, missing), total energy intake (quintiles, missing), physical activity in metabolic equivalents (METs: \8, 8 to \17.5, 17.5 to \31.5, C31.5 h/week, missing), combinations of number of live births (nulliparous, 1–2, 3?, unknown) and age at first birth (\25, 25–29, 30?, unknown), diabetes (yes, no), age at menarche (\12, 12, 13, 14? years, missing), age at menopause (\50, 50–54, 55? years), PMH use [never, former, other (ever unknown type, former vaginal estrogen, and former progesterone only)], and oral contraceptive use
323
(never, ever, missing). BMI at age 18 or weight change from age 18 to 1992 was also included in the model as indicated.
Results The women in the CPS-II Nutrition Cohort, who were born in the first half of the twentieth century, were generally lean at age 18. Only 5.4 % had a BMI that would classify them as overweight (BMI 25 to \30 kg/m2) and 1.3 % fell into the obese category (BMI C 30 kg/m2). Most of the women gained weight between age 18 and 1992. Only 15.1 % of women maintained their weight (±1.5 kg/m2) while even fewer (3.6 %) lost weight during this period. Compared to weight-stable women, those who gained weight tended to be less educated, were more likely to be never smokers and less likely to be current smokers, and were more likely to have diabetes (Table 1). Women who gained weight also were more likely to have had three or more live births at an early age, to have never used PMH, to consume more energy and less alcohol, and to exercise less. Women who gained weight also had a slightly lower BMI at age 18 and a considerably higher BMI in 1992 than weight-stable women. Compared to women with low BMI at age 18, women with higher BMI at age 18 (Supplemental Table 1) were less educated, more likely to have had diabetes, more likely to have never used PMH, and less likely to exercise. In addition, the heavier women were more likely to be nulliparous, to have never used oral contraceptives, and to be former smokers and were less likely to be never smokers. Adult BMI change (from age 18 to 1992) was statistically significantly associated with risk of endometrial cancer and risk increased in a dose-dependent manner (Table 2). Expressed as a rate, the association with adult weight change corresponded to a 67 % increased risk per BMI unit change per 10 years (RR 1.67, 95 % CI 1.55–1.80 for 1 BMI unit/10 years). Additional adjustment for BMI at age 18 increased the association slightly so that among women who gained more than 9 BMI units, risk was over six times that for weight-stable women. BMI at age 18 was also associated with higher risk of endometrial cancer in multivariable-adjusted models (Table 2). Further adjustment for adult weight change from age 18 to 1992 increased the association. Risk was low for women who were very lean at age 18 and increased in a dose-dependent fashion with women in the higher overweight and obese categories having at least a twofold higher risk. The associations of adult weight change and BMI at age 18 with risk of endometrial cancer types I and II tumors are shown in Table 3. Adult BMI change was associated with both types of tumors, although the risk estimates for type I cancers was slightly higher than for type II cancers (RR
123
324
Cancer Causes Control (2014) 25:321–328
Table 1 Age-standardizeda frequenciesb of selected characteristics of women in the Cancer Prevention Study II Nutrition Cohort, by weight change in BMI units from age 18 to 1992 Variable
Adult BMI change from age 18 to 1992 Lose [1.5 n = 1,182
Lose 1.5–gain 1.5 n = 4,987
Gain 1.6–4.0 n = 8,685
Gain 4.1–6.0 n = 6,431
Gain 6.1–9.0 n = 6,367
Gain [9.0 n = 5,405
64.0
63.7
63.5
63.5
63.4
63.0
White
98.3
98.2
97.9
98.0
97.6
96.3
Black
0.7
0.5
1.0
1.0
1.7
2.9
Other/missing
1.0
1.2
1.2
1.0
0.8
0.8
\High school
3.8
3.4
4.6
5.2
6.0
7.6
High school grade
29.7
27.4
31.8
35.7
38.3
39.2
Some college
29.8
29.5
29.5
30.6
29.3
29.5
CCollege grade
36.3
38.9
33.4
27.9
25.8
23.1
Missing
0.3
0.7
0.8
0.5
0.6
0.6
Never
43.3
51.2
53.9
55.1
56.9
59.2
Current
18.9
12.7
10.4
9.1
7.4
5.7
Former
36.4
34.7
34.4
33.9
34.3
33.6
Missing
1.4
1.4
1.3
1.8
1.4
1.5
History of diabetes
5.9
4.2
4.6
5.4
7.5
12.5
Family history of endometrial cancer
2.1
2.8
2.5
2.5
2.9
3.0
Age in 1992b Race
Education
Smoking status
Number live births, age at first birth Nulliparous
13.2
7.9
7.8
7.1
7.7
7.3
1–2, \25 years
12.7
14.1
13.8
14.2
13.0
13.5
1–2, 25–29 years
12.8
13.7
13.7
12.2
11.1
10.2
1–2, 30? years
7.6
6.5
6.1
6.1
5.6
5.4
3?, \25 years
30.8
34.8
35.3
38.2
40.9
43.5
3?, 25–29 years
17.1
17.9
17.8
17.1
16.1
14.4
3?, 30? years
3.2
2.7
3.1
2.9
2.9
2.6
2.6
2.4
2.5
2.2
2.8
3.2
Age at menarche (years)
Unknown/missing
12.8
12.8
12.8
12.8
12.8
12.6
Age at menopause (years)
50.9
50.9
50.8
50.8
50.9
51.0
Never
70.3
65.8
67.5
67.5
67.2
67.4
Ever
28.9
32.9
31.3
31.4
31.3
30.9
Unknown/missing
0.9
1.2
1.2
1.0
1.5
1.7
Never
71.6
72.0
72.7
73.6
73.3
75.0
Former
25.8
25.3
25.0
24.0
24.4
22.9
Other
2.6
2.7
2.3
2.4
2.3
2.2
BMI age 18c
25.6
21.4
20.6
20.3
20.3
20.6
BMI in 1992c
22.2
21.7
23.5
25.4
27.7
32.9
Wt D age 18 to 92c
-9.0
1.0
7.6
13.6
19.9
32.9
Energy (kcal/day)c
1,261
1,313
1,322
1,349
1,401
1,463
Exercise (METs/h)c
14.5
14.3
13.0
11.8
10.7
9.2
Alcohol (g/day)c
5.6
6.0
5.2
4.9
4.1
3.1
Oral contraceptive use
PMH use
2
BMI body mass index (kg/m ), PMH postmenopausal hormones, METs physical activity in metabolic equivalents a
Standardized to the age distribution of the women in the cohort
b
Values are presented as percentages unless otherwise noted
c
Values are presented as means
123
Cancer Causes Control (2014) 25:321–328 Table 2 Association of BMI and BMI change with endometrial cancer risk in the Cancer Prevention Study II Nutrition Cohort a
Multivariable-adjusted model includes age, alcohol use, smoking, physical activity, family history of endometrial cancer, number of live births/ age at first birth, age at menarche, age at menopause, diabetes, postmenopausal hormones use, oral contraceptive use, and energy
b
Weight variable added to the model is BMI change from age 18 to 1992 for the BMI at age 18 analyses and BMI at age 18 for the adult BMI change analysis
c
Adult BMI change is defined as BMI change (kg/m2) from age 18 to enrollment in 1992 (mean age 63.4 years)
Variable
325
Cases/person (years)
RR (95 % CI) (MV adjusted)a
RR (95 % CI) (MV ? Wt Varb)
Adult BMI changec Lose [1.5
7/12,907
1.10 (0.48–2.54)
0.79 (0.34–1.83)
Lose 1.5–gain 1.5
27/55,724
1.00 (ref)
1.00 (ref)
Gain 1.6–4.0
59/99,350
1.21 (0.77–1.90)
1.28 (0.81–2.02)
Gain 4.1–6.0
62/74,344
1.69 (1.07–2.66)
1.85 (1.17–2.91)
Gain 6.1–9.0
107/73,792
2.83 (1.85–4.33)
3.10 (2.03–4.76)
Gain [9.0
185/61,642
5.65 (3.74–8.53)
6.13 (4.05–9.27)
p trend
\0.0001
\0.0001
Per 5 BMI units change
1.81 (1.66–1.98)
1.83 (1.68–1.99)
BMI at age 18 (kg/m2) \18.5 18.5 to \22.5
65/66,600 270/237,333
0.90 (0.63–1.29) 1.05 (0.79–1.39)
0.75 (0.52, 1.07) 1.03 (0.77, 1.37)
22.5 to \25
58/49,421
1.00 (ref)
1.00 (ref)
25 to \27.5
29/15,333
1.51 (0.96–2.36)
1.52 (0.97, 2.38)
27.5 to \30
13/4,493
2.30 (1.25–4.21)
2.46 (1.34, 4.51)
C30
12/4,580
2.04 (1.09–3.81)
2.34 (1.25, 4.40)
p trend
0.0004
\0.0001
Per 5 BMI units
1.29 (1.12–1.49)
1.43 (1.24–1.65)
1.88, 95 % CI 1.69–2.09 for 5 BMI unit change for type I vs. RR 1.68, 95 % CI 1.42–2.00 for 5 BMI unit change for type II). The highest category of BMI at age 18 (BMI C 27.5) was associated with a statistically significantly increased risk of type I tumors. No significant association for BMI at age 18 was found for type II cancers, but the number of cases of heavier women at age 18 was very limited (n = 7). The combined associations of BMI at age 18 and adult weight gain with risk of endometrial cancer are shown in Table 4. Greater BMI gain was associated with greater risk, regardless of the BMI at age 18. Comparison of the hazards ratios obtained from separate analyses of the two BMI at age 18 categories showed similar results (RR 1.84, 95 % CI 1.66–2.04 per 5 BMI unit gain among women with a BMI at age 18 \22.5 and RR 1.75, 95 % CI 1.50–2.04 per 5 BMI unit gain among women with a BMI at age 18 C22.5). To investigate possible confounding from former use of estrogen-only PMH, a sensitivity analysis was conducted excluding all former users of estrogen-only PMH and former users of PMH whose type was unknown. This exclusion had no effect on the observed association (data not shown). Exclusion of women with diabetes also had no effect on these associations (data not shown).
Discussion In this large prospective study of US women, both BMI at age 18 and adult weight change were significantly
associated with risk of endometrial cancer. The 6.1-fold risk associated with the highest BMI gain of [9 units during adulthood and the twofold increased risk among women who were heavier at age 18 are generalizable to women who are never or former users of PMH. The association of adult weight change with type I endometrial cancer appeared to be stronger than with type II cancer, but the results were not significantly different. Our findings of a significant association between BMI at age 18 and risk of endometrial cancer are consistent with six [8, 17, 22–25] of the nine studies that investigated this without adjustment for BMI later in life. In the other three studies [10, 13, 19], which found no association, BMI in the normal weight range (\25 kg/m2) was included in the top category of BMI examined. This is likely to have limited the possibility of detecting a significant association. Several previous studies of early adult body size or BMI have presented the association with endometrial cancer risk after adjustment for BMI later in life [11, 12, 14–16, 21, 26, 27]. Most, but not all [14, 16], of these studies found that BMI or body size at 18 or in the early twenties was not associated with risk of endometrial cancer risk. High early adult BMI is likely to contribute causally to BMI later in life, as weight present in early adulthood is usually retained in later life. Because later life BMI is a mediator on the causal pathway between early life BMI and endometrial cancer risk, adjustment for BMI later in life would mask the association with early life BMI. Weight gain during adulthood was associated with a dramatically increased risk of endometrial cancer in this
123
326
Cancer Causes Control (2014) 25:321–328
Table 3 Association of BMI and BMI change with endometrial cancer risk by subtype in the Cancer Prevention Study II Nutrition Cohort Variable
Type 1
Type II
Cases/person.(years)
RR (95 % CI) (MV adjusted)a
Cases/person (years)
RR (95 % CI) (MV adjusted)a
Lose [1.5
4/12,907
1.11 (0.37–3.35)
2/12,907
1.00 (0.22–4.64)
Lose 1.5–gain 1.5 Gain 1.6–4.0
15/55,724 35/99,350
1.00 (ref) 1.30 (0.71–2.38)
9/55,724 19/99,350
1.00 (ref) 1.16 (0.52–2.56)
Gain 4.1–6.0
33/74,344
1.63 (0.88–3.00)
26/74,344
2.06 (0.96–4.42)
Gain 6.1–9.0
76/73,792
3.61 (2.07–6.31)
24/73,792
1.87 (0.86–4.05)
Gain [9.0
133/61,642
7.23 (4.20–12.5)
41/61,642
Adult BMI changeb
3.79 (1.81–7.94)
p trend
\0.0001
\0.0001
Per 5 BMI unit change
1.88 (1.69–2.09)
1.68 (1.42–2.00)
BMI at age 18 (kg/m2) \18.5
43/66,600
18.5 to \22.5
181/237,333
1.22 (0.85–1.77)
75/237,333
0.73 (0.45–1.18)
22.5 to \25
34/49,421
1.00 (ref)
22/49,421
1.00 (ref)
25 to \27.5c
22/15,333
1.93 (1.12–3.30)
7/24,406
0.60 (0.26–1.42)
C27.5
16/9,073
2.32 (1.27–4.22)
1.04 (0.66–1.64)
17/66,600
0.59 (0.31–1.12)
p trend
0.02
0.15
Per 5 BMI units
1.24 (1.04–1.48)
1.24 (0.93–1.65)
a
Multivariable-adjusted model includes age, alcohol use, smoking, physical activity, family history of endometrial cancer, number of live births/ age at first birth, age at menarche, age at menopause, diabetes, postmenopausal hormones use, oral contraceptive use, and energy
b
Adult weight changes defined as BMI (kg/m2) change from age 18 to enrollment in 1992 (mean age 63.4 years)
c
For type II tumors, the fourth category for BMI at age 18 is C25
Table 4 Association of weight change from age 18 to 1992 with endometrial cancer risk stratified by BMI age 18 in the Cancer Prevention Study II Nutrition Cohort Adult BMI (kg/m2) change
BMI age 18 \22.5 Cases/person (years)
BMI age 18 C22.5 RR (95 % CI)
Cases/person (years)
RR (95 % CI)
Lose [1.5
1/3,107
0.60 (0.08–4.47)
6/9,801
1.19 (0.48–2.95)
Lose 1.5–gain 1.5
22/43,307
1.00 (ref)
5/12,417
0.76 (0.29–2.01)
Gain 1.5–4.0
48/83,382
1.13 (0.68–1.87)
11/15,968
1.22 (0.59–2.53)
Gain 4.1–6.0 Gain 6.1–9.0
51/63,396 77/61,987
1.56 (0.94–2.57) 2.34 (1.45–3.77)
11/10,948 30/11,806
1.84 (0.89–3.81) 4.41 (2.52–7.70)
Gain [9.0
136/48,755
5.09 (3.22–8.06)
49/12,887
6.56 (3.91–11.0)
p for interaction
0.65
Multivariable-adjusted model includes age, alcohol use, smoking, physical activity, family history of endometrial cancer, number of live births/ age at first birth, age at menarche, age at menopause, diabetes, postmenopausal hormones use, oral contraceptive use, and energy
study. Our findings, which were not adjusted for later life BMI, are consistent with ten [9, 12, 13, 17, 19–21, 23, 24, 27] of eleven [10] other similar studies. Only two [16, 18] of the five [14, 15, 24] previous studies that presented results adjusted for later life BMI reported a positive association. Because BMI later in life is on the causal pathway between adult weight gain and endometrial cancer, adjustment for later life BMI may obscure effects of
123
adult weight gain. Thus, the present findings add to the existing evidence to indicate that adult weight change is significantly associated with risk of endometrial cancer. Endometrial cancers have distinct histologic features with different prognosis. Type I cancers are primarily endometrioid carcinomas and are associated with unopposed estrogenic stimuli [35]. Type II tumors are typically serous and clear cell adenocarcinomas that arise form polyps or
Cancer Causes Control (2014) 25:321–328
atrophic endometrial and less often express functional estrogen and progesterone receptors [35, 36]. The etiology of the rarer and more aggressive type II tumors is less well understood, although the two types share many risk factors [37]. In this analysis, BMI change was associated with higher risk of both types, while BMI at age 18 was only associated with type I tumors. The smaller number of type II tumors, particularly among heavier women at age 18, may have hampered our ability to detect this association. Adjustment for adult BMI change increased the risk associated with BMI at age 18. It is possible that this adjustment could have resulted in an overestimate of the risk associated with BMI at 18 due to conditioning on a collider [38]. While the actual hazards ratios may not be as high as indicated by the adjusted point estimates, the fact that the association between early adult BMI and higher risk of endometrial cancer persists with or without adjustment for adult BMI change provides additional support for the importance of early adult BMI as an important risk factor for endometrial cancer. This study has several strengths, including its large study population and its prospective design. This was the first study of non-PMH users to examine the independent associations of BMI at age 18 and adult weight change with endometrial cancer. In addition, the wide range of weight changes during adulthood among the women allowed us to determine risks of endometrial cancer for greater extremes of weight gain than previous studies. Limitations of this study include the fact that the information used to determine weight change was both self-reported and recalled. While this may result in some misclassification, there is no reason to expect it to be differential with respect to endometrial cancer risk. In addition, the CPS-II Nutrition Cohort included few women who were overweight or obese at age 18, which is expected for Americans born prior to 1950. However, this limited our ability to determine whether the risk of endometrial cancer associated with adult weight gain is similar for women who are overweight or obese at age 18 to that for women who are lean at this age. In summary, the results of this study indicate that both high BMI at age 18 and greater adult weight gain are associated with an increased risk of endometrial cancer. The contribution of overweight and obesity to endometrial cancer incidence and mortality is likely to increase population trends in overweight and obesity. These results suggest that risk of endometrial cancer can be minimized both by avoiding obesity in young adulthood and by maintaining a healthy weight as recommended by the American Cancer Society’s guidelines on nutrition and physical activity for cancer prevention [39]. Acknowledgments The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study II
327 (CPS-II) cohort. The authors thank the CPS-II participants and Study Management Group for their invaluable contributions to this research. The authors would also like to acknowledge the contribution to this study from central cancer registries supported through the Centers for Disease Control and Prevention National Program of Cancer Registries, and cancer registries supported by the National Cancer Institute Surveillance Epidemiology and End Results program.
References 1. Siegel R, Naishadham D, Jemal A (2012) Cancer statistics, 2012. CA Cancer J Clin 62:10–29 2. Renehan AG, Tyson M, Egger M et al (2008) Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 371:569–578 3. Calle EE, Kaaks R (2004) Overweight, obesity and cancer: epidemiological evidence and proposed mechanisms. Nat Rev Cancer 4:579–591 4. Kaaks R, Lukanova A, Kurzer MS (2002) Obesity, endogenous hormones, and endometrial cancer risk: a synthetic review. Cancer Epidemiol Biomarkers Prev 11:1531–1543 5. Pan L, Blanck HM, Sherry B et al (2012) Trends in the prevalence of extreme obesity among US preschool-aged children living in low-income families, 1998–2010. J Am Med Assoc 308:2563–2565 6. Flegal KM, Carroll MD, Kit BK et al (2012) Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999–2010. J Am Med Assoc 307:491–497 7. Crosbie EJ, Zwahlen M, Kitchener HC et al (2010) Body mass index, hormone replacement therapy, and endometrial cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 19:3119–3130 8. Blitzer PH, Blitzer EC, Rimm AA (1976) Association between teen-age obesity and cancer in 56,711 women: all cancers and endometrial cancer. Prev Med 5:20–31 9. Le Marchand L, Wilkens LR, Mi M-P (1991) Early-age body size, adult weight gain and endometrial cancer risk. Int J Cancer 48:807–811 10. Shu XO, Brinton LA, Zheng W et al (1992) Relation of obesity and body fat distribution to endometrial cancer in Shanghai, China. Cancer Res 52:3865–3870 11. Levi F, La Vecchia C, Negri E et al (1992) Body mass at different ages and subsequent endometrial cancer risk. Int J Cancer 50:567–571 12. Swanson CA, Potischman N, Wilbanks GD et al (1993) Relation of endometrial cancer risk to past and contemporary body size and body fat distribution. Cancer Epidemiol Biomarkers Prev 2:321–327 13. Olson SH, Trevisan M, Marshall JR et al (1995) Body mass index, weight gain, and risk of endometrial cancer. Nutr Cancer 23:141–149 14. Terry P, Baron JA, Weiderpass E et al (1999) Lifestyle and endometrial cancer risk: a cohort study from the Swedish Twin Registry. Int J Cancer 82:38–42 15. Weiderpass E, Persson I, Adami HO et al (2000) Body size in different periods of life, diabetes mellitus, hypertension, and risk of postmenopausal endometrial cancer (Sweden). Cancer Causes Control 11:185–192 16. Jonsson F, Wolk A, Pedersen NL et al (2003) Obesity and hormone-dependent tumors: cohort and co-twin control studies based on the Swedish Twin Registry. Int J Cancer 106:594–599 17. Schouten LJ, Goldbohm RA, van den Brandt PA (2004) Anthropometry, physical activity, and endometrial cancer risk: results from the Netherlands Cohort Study. J Natl Cancer Inst 96:1635–1638
123
328 18. Trentham-Dietz A, Nichols HB, Hampton JM et al (2006) Weight change and risk of endometrial cancer. Int J Epidemiol 35: 151–156 19. Xu WH, Xiang YB, Zheng W et al (2006) Weight history and risk of endometrial cancer among Chinese women. Int J Epidemiol 35:159–166 20. Friedenreich C, Cust A, Lahmann PH et al (2007) Anthropometric factors and risk of endometrial cancer: the European prospective investigation into cancer and nutrition. Cancer Causes Control 18:399–413 21. Chang SC, Lacey JV Jr, Brinton LA et al (2007) Lifetime weight history and endometrial cancer risk by type of menopausal hormone use in the NIH-AARP diet and health study. Cancer Epidemiol Biomarkers Prev 16:723–730 22. Thomas CC, Wingo PA, Dolan MS et al (2009) Endometrial cancer risk among younger, overweight women. Obstet Gynecol 114:22–27 23. Park SL, Goodman MT, Zhang ZF et al (2010) Body size, adult BMI gain and endometrial cancer risk: the multiethnic cohort. Int J Cancer 126:490–499 24. Hosono S, Matsuo K, Hirose K et al (2011) Weight gain during adulthood and body weight at age 20 are associated with the risk of endometrial cancer in Japanese women. J Epidemiol 21:466–473 25. Lu L, Risch H, Irwin ML et al (2011) Long-term overweight and weight gain in early adulthood in association with risk of endometrial cancer. Int J Cancer 129:1237–1243 26. Yang TY, Cairns BJ, Allen N et al (2012) Postmenopausal endometrial cancer risk and body size in early life and middle age: prospective cohort study. Br J Cancer 107:169–175 27. Nagle CM, Marquart L, Bain CJ et al (2013) Impact of weight change and weight cycling on risk of different subtypes of endometrial cancer. Eur J Cancer 49:2717–2726 28. McCullough ML, Patel AV, Patel R et al (2008) Body mass and endometrial cancer risk by hormone replacement therapy and cancer subtype. Cancer Epidemiol Biomarkers Prev 17:73–79
123
Cancer Causes Control (2014) 25:321–328 29. Rossouw JE, Anderson GL, Prentice RL et al (2002) Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women’s Health Initiative randomized controlled trial. J Am Med Assoc 288:321–333 30. Sprague BL, Trentham-Dietz A, Cronin KA (2012) A sustained decline in postmenopausal hormone use: results from the National Health and Nutrition Examination Survey, 1999–2010. Obstet Gynecol 120:595–603 31. Calle EE, Rodriguez C, Jacobs EJ et al (2002) The American Cancer Society Cancer Prevention Study II Nutrition Cohort. Cancer 94:2490–2501 32. Bergmann MM, Calle EE, Mervis CA et al (1998) Validity of selfreported cancer in a prospective cohort study in comparison with data form state cancer registries. Am J Epidemiol 147:556–562 33. Prat J (2004) Prognostic parameters of endometrial carcinoma. Hum Pathol 35:649–662 34. Cox D (1972) Regression models and life tables. J R Stat Soc B 34:187–202 35. Sherman ME (2000) Theories of endometrial carcinogenesis: a multidisciplinary approach. Mod Path 13:295–308 36. Emons G, Fleckenstein G, Hinney B et al (2000) Hormonal interactions in endometrial cancer. Endocr Relat Cancer 7:227–242 37. Setiawan VW, Yang HP, Pike MC et al (2013) Type I and II endometrial cancers: have they different risk factors? J Clin Oncol 31:2607–2618 38. Cole SR, Platt RW, Schisterman EF et al (2010) Illustrating bias due to conditioning on a collider. Int J Epidemiol 39:417–420 39. Kushi LH, Doyle C, McCullough M et al (2012) American Cancer Society Guidelines on nutrition and physical activity for cancer prevention: reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin 62:30–67
ORIGINAL PAPER
Body weight in early adulthood, adult weight gain, and risk of endometrial cancer in women not using postmenopausal hormones Victoria L. Stevens • Eric J. Jacobs • Alpa V. Patel • Juzhong Sun • Susan M. Gapstur Marjorie L. McCullough
•
Received: 18 September 2013 / Accepted: 20 December 2013 / Published online: 1 January 2014 Ó Springer Science+Business Media Dordrecht 2013
Abstract Purpose High body mass index (BMI) measured in middle age or later is an established risk factor for endometrial cancer. However, whether BMI measured in early adulthood and adult weight change are associated with endometrial cancer risk is less clear, particularly among nonusers of postmenopausal hormones (PMH). Methods These associations were investigated among women in the Cancer Prevention Study II Nutrition Cohort. Women taking PMH (n = 11,624, 12 % of all women) were excluded, and the analysis was limited to 33,057 postmenopausal women who did not take PMH. Between enrollment in 1992/1993 and 30 June 2009, 447 women were diagnosed with endometrial cancer. Cox proportional hazards regression was used to calculate hazard rate ratios (RR) and corresponding 95 % confidence intervals (CI) for the association of BMI at age 18, calculated from recalled weight, and weight change between age 18 and 1992, with endometrial cancer incidence. Results BMI at age 18 was associated with higher risk of endometrial cancer in multivariable models adjusted for other risk factors (RR 1.29, 95 % CI 1.12–1.49 per 5 BMI units). Similarly, adult weight change was associated with higher risk of endometrial cancer (RR 1.81, 95 % CI 1.66–1.98 per 5 BMI unit change) in multivariable models adjusted for other risk factors.
Electronic supplementary material The online version of this article (doi:10.1007/s10552-013-0333-7) contains supplementary material, which is available to authorized users. V. L. Stevens (&) E. J. Jacobs A. V. Patel J. Sun S. M. Gapstur M. L. McCullough Epidemiology Research Program, American Cancer Society, 250 Williams St, NW, Atlanta, GA 30329, USA e-mail: [email protected]
Conclusions High BMI at age 18 and greater adult weight gain were strongly associated with risk of endometrial cancer. These results underscore the importance of both avoiding overweight/obesity in young adulthood and preventing weight gain thereafter to minimize risk of this cancer. Keywords Endometrial cancer Obesity Weight change Postmenopausal
Introduction Endometrial cancer is the most common gynecologic cancer and fourth most common cancer among women in the USA [1]. Obesity is an established risk factor for this cancer; it is estimated that risk increases about 59 % for every 5 unit increase in body mass index (BMI, kg/m2) [2], and overweight and obesity are responsible for 57 % of all cases of endometrial cancer in the USA [3]. Like most risk factors for this cancer, obesity increases exposure to estrogen unopposed by progesterone in both premenopausal and postmenopausal women [4]. With the dramatic increases in the prevalence of overweight and obesity in both children [5] and adults [6] in recent decades, a clear understanding of the consequences of excess bodyweight early in life and weight gain as an adult is essential knowledge for physicians to appropriately advise their patients as to how to minimize cancer risk. The studies which established obesity as a risk factor for endometrial cancer measured BMI at or near the time of diagnosis [7], which is usually after the age of 50. Despite numerous epidemiologic studies [8–27], the associations of BMI in early adulthood and adult weight change with endometrial cancer are not well defined. Many of the
123
322
previous studies only investigated one of these exposures [8, 9, 11, 18, 20, 22, 25], and several of the studies of both BMI in early adulthood and adult weight change [12, 14– 16, 21, 27] adjusted for later life BMI in their statistical model, which may obscure the contribution of these factors to increased risk of endometrial cancer. Five [13, 17, 19, 23, 24] of the six [10] studies that did not adjust for later life BMI reported statistically significant positive associations of adult weight change with endometrial cancer risk, but the findings for early adulthood BMI were mixed with three studies reporting a positive association [17, 23, 24] and three reporting null results [10, 13, 19]. An additional consideration for understanding the association of BMI at early adulthood and adult weight change with endometrial cancer risk is that the use of postmenopausal hormone therapy (PMH) significantly attenuates the association between obesity and endometrial cancer [20, 21, 28]. With one exception [26], all previous studies of BMI in early adulthood and/or adult weight change with endometrial cancer risk [8–25, 27] included both users and nonusers of PMH. Since publication of results from the Women’s Health Initiative (WHI) [29], the prevalence of PMH had been declining steadily [30]. Because future populations of women will include few long-term PMH users, it is important to examine the relative risks of early adult BMI and adult weight change associated with endometrial cancer in non-PMH users. We examined the associations of BMI at age 18 and adult weight change with endometrial cancer in the American Cancer Society’s Cancer Prevention Study II (CPS-II) Nutrition Cohort, a large prospective study of older adults. These analyses were conducted among women not using PMH so that the results would be applicable to women currently seeking advice on minimizing cancer risk.
Materials and methods Study population Subjects for this study were female participants from the CPS-II Nutrition Cohort, a prospective study of cancer incidence and mortality among 86,404 men and 97,785 women. The Nutrition Cohort, which is described in detail elsewhere [31], was initiated in 1992 as a subcohort of CPS-II, a prospective study of cancer mortality begun in 1982 involving approximately 1.2 million Americans. Nutrition Cohort participants were recruited from CPS-II members who resided in 21 states and were predominantly between the ages of 50 and 74 years. Participants completed a self-administered questionnaire at baseline that included demographic, medical, anthropometric, dietary,
123
Cancer Causes Control (2014) 25:321–328
and lifestyle information. Follow-up questionnaires were sent to all living Nutrition Cohort members in 1997 and every two years thereafter to update exposure information and to ascertain newly diagnosed cancers. The response rate on all of the follow-up questionnaires, among those cohort participants who were mailed surveys, was at least 87 %. Follow-up was from the date of receipt of the baseline survey in 1992/1993 until the date of diagnosis of endometrial cancer, death, date of the last survey returned, or 30 June 2009 whichever came first. All aspects of the CPS-II Nutrition Cohort study are approved by the Emory University Institutional Review Board. Women were excluded from this analysis if they were lost to follow-up (n = 3,111), reported any prevalent cancer other than non-melanoma skin cancer in either 1982 (n = 6,413) or 1992 (n = 6,681), or self-reported endometrial cancer on their first follow-up survey which could not be verified (n = 45). Women who reported endometrial cancer during later follow-up that could not be verified were censored on the date of their last cancer-free survey. Also excluded were women with missing BMI at age 18 or in 1992 and who reported a BMI \ 10 kg/m2 at age 18 or BMI \ 18.5 kg/m2 in 1992 (n = 3,658). Women who were premenopausal or whose menopausal status was unknown (n = 4,124) and who reported that their uterus had been removed or had unknown uterine status at baseline (n = 28,943) were also excluded. Finally, because postmenopausal hormone (PMH) use is known to modify the association between BMI and risk of endometrial cancer, all women who reported current use of PMH or for whom use of PMH was unknown in 1992 (n = 11,753) were also excluded. Additionally, women who started taking PMH after baseline were censored on the date of receipt of the questionnaire on which this use was reported. Women who did not return two consecutive surveys but returned a later survey were censored 4 months after the second missed survey was mailed, and women who underwent hysterectomy after enrollment were censored at the date of receipt of the questionnaire on which this procedure was reported. The final analytic cohort for this study included 33,057 women. Endometrial cancer cases After all exclusions, we identified and verified 447 cases of endometrial cancer diagnosed between enrollment in 1992/1993 and 30 June 2009 for this analysis. These cases (ICD-0 code 54.1) were initially identified through selfreport on the follow-up questionnaires and subsequently verified through medical records (n = 286) or linkage with state cancer registries (n = 161) [32]. Endometrial cancer cases were further subtyped based on histology into endometrioid (type I) cancers (n = 296)
Cancer Causes Control (2014) 25:321–328
and high grade, poor prognosis (type II) cancers (n = 121). Type II cancers included serous papillary, clear cell, or squamous cell histology (n = 34), and any additional International Federation of Gynecology and Obstetrics grade three or four cancers (n = 87) [33]. Cases with missing grade (n = 30) were excluded from subtype analyses. BMI and adult weight change Women self-reported their current weight and weight at age 18 on the Nutrition Cohort baseline questionnaire in 1992. These weights were used to calculate BMI by dividing by the square of their height, which was ascertained in 1982. Weight change between age 18 and 1992 was determined from the difference between two values given for each of these times and is expressed as BMI change. The rate of BMI change from age 18 to 1992 was determined by dividing the BMI change by 10 years. Statistical analyses All statistical analyses were conducted with SAS v9.3. Cox proportional hazards regression [34] was used to calculate hazard rate ratios (RR) and corresponding 95 % confidence intervals (CI) for the association of BMI and weight change, modeled as categorical and continuous variables, with endometrial cancer incidence. p values for linear trend were determined using continuous variables for BMI and weight change. The Cox proportional hazards assumption was tested by modeling a multiplicative interaction term between BMI at 18 or weight change and time. No violations were found. Interactions between BMI at age 18 and weight change were tested using the likelihood ratio test to determine whether the association of weight gain differed between women of different body sizes at early adulthood. All Cox models were adjusted for age by stratifying on exact year of age in 1992. Additional covariates included in the multivariable-adjusted models were family history of endometrial cancer (yes, no), alcohol consumption (not current drinker, \1 drink/week, 1–6 drinks/week, 1 drink/ day, C2 drinks/day, missing), smoking status (never, former \20 years, former C20 years, current \40 years, current C40 years, ever/unclassifiable, missing), total energy intake (quintiles, missing), physical activity in metabolic equivalents (METs: \8, 8 to \17.5, 17.5 to \31.5, C31.5 h/week, missing), combinations of number of live births (nulliparous, 1–2, 3?, unknown) and age at first birth (\25, 25–29, 30?, unknown), diabetes (yes, no), age at menarche (\12, 12, 13, 14? years, missing), age at menopause (\50, 50–54, 55? years), PMH use [never, former, other (ever unknown type, former vaginal estrogen, and former progesterone only)], and oral contraceptive use
323
(never, ever, missing). BMI at age 18 or weight change from age 18 to 1992 was also included in the model as indicated.
Results The women in the CPS-II Nutrition Cohort, who were born in the first half of the twentieth century, were generally lean at age 18. Only 5.4 % had a BMI that would classify them as overweight (BMI 25 to \30 kg/m2) and 1.3 % fell into the obese category (BMI C 30 kg/m2). Most of the women gained weight between age 18 and 1992. Only 15.1 % of women maintained their weight (±1.5 kg/m2) while even fewer (3.6 %) lost weight during this period. Compared to weight-stable women, those who gained weight tended to be less educated, were more likely to be never smokers and less likely to be current smokers, and were more likely to have diabetes (Table 1). Women who gained weight also were more likely to have had three or more live births at an early age, to have never used PMH, to consume more energy and less alcohol, and to exercise less. Women who gained weight also had a slightly lower BMI at age 18 and a considerably higher BMI in 1992 than weight-stable women. Compared to women with low BMI at age 18, women with higher BMI at age 18 (Supplemental Table 1) were less educated, more likely to have had diabetes, more likely to have never used PMH, and less likely to exercise. In addition, the heavier women were more likely to be nulliparous, to have never used oral contraceptives, and to be former smokers and were less likely to be never smokers. Adult BMI change (from age 18 to 1992) was statistically significantly associated with risk of endometrial cancer and risk increased in a dose-dependent manner (Table 2). Expressed as a rate, the association with adult weight change corresponded to a 67 % increased risk per BMI unit change per 10 years (RR 1.67, 95 % CI 1.55–1.80 for 1 BMI unit/10 years). Additional adjustment for BMI at age 18 increased the association slightly so that among women who gained more than 9 BMI units, risk was over six times that for weight-stable women. BMI at age 18 was also associated with higher risk of endometrial cancer in multivariable-adjusted models (Table 2). Further adjustment for adult weight change from age 18 to 1992 increased the association. Risk was low for women who were very lean at age 18 and increased in a dose-dependent fashion with women in the higher overweight and obese categories having at least a twofold higher risk. The associations of adult weight change and BMI at age 18 with risk of endometrial cancer types I and II tumors are shown in Table 3. Adult BMI change was associated with both types of tumors, although the risk estimates for type I cancers was slightly higher than for type II cancers (RR
123
324
Cancer Causes Control (2014) 25:321–328
Table 1 Age-standardizeda frequenciesb of selected characteristics of women in the Cancer Prevention Study II Nutrition Cohort, by weight change in BMI units from age 18 to 1992 Variable
Adult BMI change from age 18 to 1992 Lose [1.5 n = 1,182
Lose 1.5–gain 1.5 n = 4,987
Gain 1.6–4.0 n = 8,685
Gain 4.1–6.0 n = 6,431
Gain 6.1–9.0 n = 6,367
Gain [9.0 n = 5,405
64.0
63.7
63.5
63.5
63.4
63.0
White
98.3
98.2
97.9
98.0
97.6
96.3
Black
0.7
0.5
1.0
1.0
1.7
2.9
Other/missing
1.0
1.2
1.2
1.0
0.8
0.8
\High school
3.8
3.4
4.6
5.2
6.0
7.6
High school grade
29.7
27.4
31.8
35.7
38.3
39.2
Some college
29.8
29.5
29.5
30.6
29.3
29.5
CCollege grade
36.3
38.9
33.4
27.9
25.8
23.1
Missing
0.3
0.7
0.8
0.5
0.6
0.6
Never
43.3
51.2
53.9
55.1
56.9
59.2
Current
18.9
12.7
10.4
9.1
7.4
5.7
Former
36.4
34.7
34.4
33.9
34.3
33.6
Missing
1.4
1.4
1.3
1.8
1.4
1.5
History of diabetes
5.9
4.2
4.6
5.4
7.5
12.5
Family history of endometrial cancer
2.1
2.8
2.5
2.5
2.9
3.0
Age in 1992b Race
Education
Smoking status
Number live births, age at first birth Nulliparous
13.2
7.9
7.8
7.1
7.7
7.3
1–2, \25 years
12.7
14.1
13.8
14.2
13.0
13.5
1–2, 25–29 years
12.8
13.7
13.7
12.2
11.1
10.2
1–2, 30? years
7.6
6.5
6.1
6.1
5.6
5.4
3?, \25 years
30.8
34.8
35.3
38.2
40.9
43.5
3?, 25–29 years
17.1
17.9
17.8
17.1
16.1
14.4
3?, 30? years
3.2
2.7
3.1
2.9
2.9
2.6
2.6
2.4
2.5
2.2
2.8
3.2
Age at menarche (years)
Unknown/missing
12.8
12.8
12.8
12.8
12.8
12.6
Age at menopause (years)
50.9
50.9
50.8
50.8
50.9
51.0
Never
70.3
65.8
67.5
67.5
67.2
67.4
Ever
28.9
32.9
31.3
31.4
31.3
30.9
Unknown/missing
0.9
1.2
1.2
1.0
1.5
1.7
Never
71.6
72.0
72.7
73.6
73.3
75.0
Former
25.8
25.3
25.0
24.0
24.4
22.9
Other
2.6
2.7
2.3
2.4
2.3
2.2
BMI age 18c
25.6
21.4
20.6
20.3
20.3
20.6
BMI in 1992c
22.2
21.7
23.5
25.4
27.7
32.9
Wt D age 18 to 92c
-9.0
1.0
7.6
13.6
19.9
32.9
Energy (kcal/day)c
1,261
1,313
1,322
1,349
1,401
1,463
Exercise (METs/h)c
14.5
14.3
13.0
11.8
10.7
9.2
Alcohol (g/day)c
5.6
6.0
5.2
4.9
4.1
3.1
Oral contraceptive use
PMH use
2
BMI body mass index (kg/m ), PMH postmenopausal hormones, METs physical activity in metabolic equivalents a
Standardized to the age distribution of the women in the cohort
b
Values are presented as percentages unless otherwise noted
c
Values are presented as means
123
Cancer Causes Control (2014) 25:321–328 Table 2 Association of BMI and BMI change with endometrial cancer risk in the Cancer Prevention Study II Nutrition Cohort a
Multivariable-adjusted model includes age, alcohol use, smoking, physical activity, family history of endometrial cancer, number of live births/ age at first birth, age at menarche, age at menopause, diabetes, postmenopausal hormones use, oral contraceptive use, and energy
b
Weight variable added to the model is BMI change from age 18 to 1992 for the BMI at age 18 analyses and BMI at age 18 for the adult BMI change analysis
c
Adult BMI change is defined as BMI change (kg/m2) from age 18 to enrollment in 1992 (mean age 63.4 years)
Variable
325
Cases/person (years)
RR (95 % CI) (MV adjusted)a
RR (95 % CI) (MV ? Wt Varb)
Adult BMI changec Lose [1.5
7/12,907
1.10 (0.48–2.54)
0.79 (0.34–1.83)
Lose 1.5–gain 1.5
27/55,724
1.00 (ref)
1.00 (ref)
Gain 1.6–4.0
59/99,350
1.21 (0.77–1.90)
1.28 (0.81–2.02)
Gain 4.1–6.0
62/74,344
1.69 (1.07–2.66)
1.85 (1.17–2.91)
Gain 6.1–9.0
107/73,792
2.83 (1.85–4.33)
3.10 (2.03–4.76)
Gain [9.0
185/61,642
5.65 (3.74–8.53)
6.13 (4.05–9.27)
p trend
\0.0001
\0.0001
Per 5 BMI units change
1.81 (1.66–1.98)
1.83 (1.68–1.99)
BMI at age 18 (kg/m2) \18.5 18.5 to \22.5
65/66,600 270/237,333
0.90 (0.63–1.29) 1.05 (0.79–1.39)
0.75 (0.52, 1.07) 1.03 (0.77, 1.37)
22.5 to \25
58/49,421
1.00 (ref)
1.00 (ref)
25 to \27.5
29/15,333
1.51 (0.96–2.36)
1.52 (0.97, 2.38)
27.5 to \30
13/4,493
2.30 (1.25–4.21)
2.46 (1.34, 4.51)
C30
12/4,580
2.04 (1.09–3.81)
2.34 (1.25, 4.40)
p trend
0.0004
\0.0001
Per 5 BMI units
1.29 (1.12–1.49)
1.43 (1.24–1.65)
1.88, 95 % CI 1.69–2.09 for 5 BMI unit change for type I vs. RR 1.68, 95 % CI 1.42–2.00 for 5 BMI unit change for type II). The highest category of BMI at age 18 (BMI C 27.5) was associated with a statistically significantly increased risk of type I tumors. No significant association for BMI at age 18 was found for type II cancers, but the number of cases of heavier women at age 18 was very limited (n = 7). The combined associations of BMI at age 18 and adult weight gain with risk of endometrial cancer are shown in Table 4. Greater BMI gain was associated with greater risk, regardless of the BMI at age 18. Comparison of the hazards ratios obtained from separate analyses of the two BMI at age 18 categories showed similar results (RR 1.84, 95 % CI 1.66–2.04 per 5 BMI unit gain among women with a BMI at age 18 \22.5 and RR 1.75, 95 % CI 1.50–2.04 per 5 BMI unit gain among women with a BMI at age 18 C22.5). To investigate possible confounding from former use of estrogen-only PMH, a sensitivity analysis was conducted excluding all former users of estrogen-only PMH and former users of PMH whose type was unknown. This exclusion had no effect on the observed association (data not shown). Exclusion of women with diabetes also had no effect on these associations (data not shown).
Discussion In this large prospective study of US women, both BMI at age 18 and adult weight change were significantly
associated with risk of endometrial cancer. The 6.1-fold risk associated with the highest BMI gain of [9 units during adulthood and the twofold increased risk among women who were heavier at age 18 are generalizable to women who are never or former users of PMH. The association of adult weight change with type I endometrial cancer appeared to be stronger than with type II cancer, but the results were not significantly different. Our findings of a significant association between BMI at age 18 and risk of endometrial cancer are consistent with six [8, 17, 22–25] of the nine studies that investigated this without adjustment for BMI later in life. In the other three studies [10, 13, 19], which found no association, BMI in the normal weight range (\25 kg/m2) was included in the top category of BMI examined. This is likely to have limited the possibility of detecting a significant association. Several previous studies of early adult body size or BMI have presented the association with endometrial cancer risk after adjustment for BMI later in life [11, 12, 14–16, 21, 26, 27]. Most, but not all [14, 16], of these studies found that BMI or body size at 18 or in the early twenties was not associated with risk of endometrial cancer risk. High early adult BMI is likely to contribute causally to BMI later in life, as weight present in early adulthood is usually retained in later life. Because later life BMI is a mediator on the causal pathway between early life BMI and endometrial cancer risk, adjustment for BMI later in life would mask the association with early life BMI. Weight gain during adulthood was associated with a dramatically increased risk of endometrial cancer in this
123
326
Cancer Causes Control (2014) 25:321–328
Table 3 Association of BMI and BMI change with endometrial cancer risk by subtype in the Cancer Prevention Study II Nutrition Cohort Variable
Type 1
Type II
Cases/person.(years)
RR (95 % CI) (MV adjusted)a
Cases/person (years)
RR (95 % CI) (MV adjusted)a
Lose [1.5
4/12,907
1.11 (0.37–3.35)
2/12,907
1.00 (0.22–4.64)
Lose 1.5–gain 1.5 Gain 1.6–4.0
15/55,724 35/99,350
1.00 (ref) 1.30 (0.71–2.38)
9/55,724 19/99,350
1.00 (ref) 1.16 (0.52–2.56)
Gain 4.1–6.0
33/74,344
1.63 (0.88–3.00)
26/74,344
2.06 (0.96–4.42)
Gain 6.1–9.0
76/73,792
3.61 (2.07–6.31)
24/73,792
1.87 (0.86–4.05)
Gain [9.0
133/61,642
7.23 (4.20–12.5)
41/61,642
Adult BMI changeb
3.79 (1.81–7.94)
p trend
\0.0001
\0.0001
Per 5 BMI unit change
1.88 (1.69–2.09)
1.68 (1.42–2.00)
BMI at age 18 (kg/m2) \18.5
43/66,600
18.5 to \22.5
181/237,333
1.22 (0.85–1.77)
75/237,333
0.73 (0.45–1.18)
22.5 to \25
34/49,421
1.00 (ref)
22/49,421
1.00 (ref)
25 to \27.5c
22/15,333
1.93 (1.12–3.30)
7/24,406
0.60 (0.26–1.42)
C27.5
16/9,073
2.32 (1.27–4.22)
1.04 (0.66–1.64)
17/66,600
0.59 (0.31–1.12)
p trend
0.02
0.15
Per 5 BMI units
1.24 (1.04–1.48)
1.24 (0.93–1.65)
a
Multivariable-adjusted model includes age, alcohol use, smoking, physical activity, family history of endometrial cancer, number of live births/ age at first birth, age at menarche, age at menopause, diabetes, postmenopausal hormones use, oral contraceptive use, and energy
b
Adult weight changes defined as BMI (kg/m2) change from age 18 to enrollment in 1992 (mean age 63.4 years)
c
For type II tumors, the fourth category for BMI at age 18 is C25
Table 4 Association of weight change from age 18 to 1992 with endometrial cancer risk stratified by BMI age 18 in the Cancer Prevention Study II Nutrition Cohort Adult BMI (kg/m2) change
BMI age 18 \22.5 Cases/person (years)
BMI age 18 C22.5 RR (95 % CI)
Cases/person (years)
RR (95 % CI)
Lose [1.5
1/3,107
0.60 (0.08–4.47)
6/9,801
1.19 (0.48–2.95)
Lose 1.5–gain 1.5
22/43,307
1.00 (ref)
5/12,417
0.76 (0.29–2.01)
Gain 1.5–4.0
48/83,382
1.13 (0.68–1.87)
11/15,968
1.22 (0.59–2.53)
Gain 4.1–6.0 Gain 6.1–9.0
51/63,396 77/61,987
1.56 (0.94–2.57) 2.34 (1.45–3.77)
11/10,948 30/11,806
1.84 (0.89–3.81) 4.41 (2.52–7.70)
Gain [9.0
136/48,755
5.09 (3.22–8.06)
49/12,887
6.56 (3.91–11.0)
p for interaction
0.65
Multivariable-adjusted model includes age, alcohol use, smoking, physical activity, family history of endometrial cancer, number of live births/ age at first birth, age at menarche, age at menopause, diabetes, postmenopausal hormones use, oral contraceptive use, and energy
study. Our findings, which were not adjusted for later life BMI, are consistent with ten [9, 12, 13, 17, 19–21, 23, 24, 27] of eleven [10] other similar studies. Only two [16, 18] of the five [14, 15, 24] previous studies that presented results adjusted for later life BMI reported a positive association. Because BMI later in life is on the causal pathway between adult weight gain and endometrial cancer, adjustment for later life BMI may obscure effects of
123
adult weight gain. Thus, the present findings add to the existing evidence to indicate that adult weight change is significantly associated with risk of endometrial cancer. Endometrial cancers have distinct histologic features with different prognosis. Type I cancers are primarily endometrioid carcinomas and are associated with unopposed estrogenic stimuli [35]. Type II tumors are typically serous and clear cell adenocarcinomas that arise form polyps or
Cancer Causes Control (2014) 25:321–328
atrophic endometrial and less often express functional estrogen and progesterone receptors [35, 36]. The etiology of the rarer and more aggressive type II tumors is less well understood, although the two types share many risk factors [37]. In this analysis, BMI change was associated with higher risk of both types, while BMI at age 18 was only associated with type I tumors. The smaller number of type II tumors, particularly among heavier women at age 18, may have hampered our ability to detect this association. Adjustment for adult BMI change increased the risk associated with BMI at age 18. It is possible that this adjustment could have resulted in an overestimate of the risk associated with BMI at 18 due to conditioning on a collider [38]. While the actual hazards ratios may not be as high as indicated by the adjusted point estimates, the fact that the association between early adult BMI and higher risk of endometrial cancer persists with or without adjustment for adult BMI change provides additional support for the importance of early adult BMI as an important risk factor for endometrial cancer. This study has several strengths, including its large study population and its prospective design. This was the first study of non-PMH users to examine the independent associations of BMI at age 18 and adult weight change with endometrial cancer. In addition, the wide range of weight changes during adulthood among the women allowed us to determine risks of endometrial cancer for greater extremes of weight gain than previous studies. Limitations of this study include the fact that the information used to determine weight change was both self-reported and recalled. While this may result in some misclassification, there is no reason to expect it to be differential with respect to endometrial cancer risk. In addition, the CPS-II Nutrition Cohort included few women who were overweight or obese at age 18, which is expected for Americans born prior to 1950. However, this limited our ability to determine whether the risk of endometrial cancer associated with adult weight gain is similar for women who are overweight or obese at age 18 to that for women who are lean at this age. In summary, the results of this study indicate that both high BMI at age 18 and greater adult weight gain are associated with an increased risk of endometrial cancer. The contribution of overweight and obesity to endometrial cancer incidence and mortality is likely to increase population trends in overweight and obesity. These results suggest that risk of endometrial cancer can be minimized both by avoiding obesity in young adulthood and by maintaining a healthy weight as recommended by the American Cancer Society’s guidelines on nutrition and physical activity for cancer prevention [39]. Acknowledgments The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study II
327 (CPS-II) cohort. The authors thank the CPS-II participants and Study Management Group for their invaluable contributions to this research. The authors would also like to acknowledge the contribution to this study from central cancer registries supported through the Centers for Disease Control and Prevention National Program of Cancer Registries, and cancer registries supported by the National Cancer Institute Surveillance Epidemiology and End Results program.
References 1. Siegel R, Naishadham D, Jemal A (2012) Cancer statistics, 2012. CA Cancer J Clin 62:10–29 2. Renehan AG, Tyson M, Egger M et al (2008) Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 371:569–578 3. Calle EE, Kaaks R (2004) Overweight, obesity and cancer: epidemiological evidence and proposed mechanisms. Nat Rev Cancer 4:579–591 4. Kaaks R, Lukanova A, Kurzer MS (2002) Obesity, endogenous hormones, and endometrial cancer risk: a synthetic review. Cancer Epidemiol Biomarkers Prev 11:1531–1543 5. Pan L, Blanck HM, Sherry B et al (2012) Trends in the prevalence of extreme obesity among US preschool-aged children living in low-income families, 1998–2010. J Am Med Assoc 308:2563–2565 6. Flegal KM, Carroll MD, Kit BK et al (2012) Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999–2010. J Am Med Assoc 307:491–497 7. Crosbie EJ, Zwahlen M, Kitchener HC et al (2010) Body mass index, hormone replacement therapy, and endometrial cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 19:3119–3130 8. Blitzer PH, Blitzer EC, Rimm AA (1976) Association between teen-age obesity and cancer in 56,711 women: all cancers and endometrial cancer. Prev Med 5:20–31 9. Le Marchand L, Wilkens LR, Mi M-P (1991) Early-age body size, adult weight gain and endometrial cancer risk. Int J Cancer 48:807–811 10. Shu XO, Brinton LA, Zheng W et al (1992) Relation of obesity and body fat distribution to endometrial cancer in Shanghai, China. Cancer Res 52:3865–3870 11. Levi F, La Vecchia C, Negri E et al (1992) Body mass at different ages and subsequent endometrial cancer risk. Int J Cancer 50:567–571 12. Swanson CA, Potischman N, Wilbanks GD et al (1993) Relation of endometrial cancer risk to past and contemporary body size and body fat distribution. Cancer Epidemiol Biomarkers Prev 2:321–327 13. Olson SH, Trevisan M, Marshall JR et al (1995) Body mass index, weight gain, and risk of endometrial cancer. Nutr Cancer 23:141–149 14. Terry P, Baron JA, Weiderpass E et al (1999) Lifestyle and endometrial cancer risk: a cohort study from the Swedish Twin Registry. Int J Cancer 82:38–42 15. Weiderpass E, Persson I, Adami HO et al (2000) Body size in different periods of life, diabetes mellitus, hypertension, and risk of postmenopausal endometrial cancer (Sweden). Cancer Causes Control 11:185–192 16. Jonsson F, Wolk A, Pedersen NL et al (2003) Obesity and hormone-dependent tumors: cohort and co-twin control studies based on the Swedish Twin Registry. Int J Cancer 106:594–599 17. Schouten LJ, Goldbohm RA, van den Brandt PA (2004) Anthropometry, physical activity, and endometrial cancer risk: results from the Netherlands Cohort Study. J Natl Cancer Inst 96:1635–1638
123
328 18. Trentham-Dietz A, Nichols HB, Hampton JM et al (2006) Weight change and risk of endometrial cancer. Int J Epidemiol 35: 151–156 19. Xu WH, Xiang YB, Zheng W et al (2006) Weight history and risk of endometrial cancer among Chinese women. Int J Epidemiol 35:159–166 20. Friedenreich C, Cust A, Lahmann PH et al (2007) Anthropometric factors and risk of endometrial cancer: the European prospective investigation into cancer and nutrition. Cancer Causes Control 18:399–413 21. Chang SC, Lacey JV Jr, Brinton LA et al (2007) Lifetime weight history and endometrial cancer risk by type of menopausal hormone use in the NIH-AARP diet and health study. Cancer Epidemiol Biomarkers Prev 16:723–730 22. Thomas CC, Wingo PA, Dolan MS et al (2009) Endometrial cancer risk among younger, overweight women. Obstet Gynecol 114:22–27 23. Park SL, Goodman MT, Zhang ZF et al (2010) Body size, adult BMI gain and endometrial cancer risk: the multiethnic cohort. Int J Cancer 126:490–499 24. Hosono S, Matsuo K, Hirose K et al (2011) Weight gain during adulthood and body weight at age 20 are associated with the risk of endometrial cancer in Japanese women. J Epidemiol 21:466–473 25. Lu L, Risch H, Irwin ML et al (2011) Long-term overweight and weight gain in early adulthood in association with risk of endometrial cancer. Int J Cancer 129:1237–1243 26. Yang TY, Cairns BJ, Allen N et al (2012) Postmenopausal endometrial cancer risk and body size in early life and middle age: prospective cohort study. Br J Cancer 107:169–175 27. Nagle CM, Marquart L, Bain CJ et al (2013) Impact of weight change and weight cycling on risk of different subtypes of endometrial cancer. Eur J Cancer 49:2717–2726 28. McCullough ML, Patel AV, Patel R et al (2008) Body mass and endometrial cancer risk by hormone replacement therapy and cancer subtype. Cancer Epidemiol Biomarkers Prev 17:73–79
123
Cancer Causes Control (2014) 25:321–328 29. Rossouw JE, Anderson GL, Prentice RL et al (2002) Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women’s Health Initiative randomized controlled trial. J Am Med Assoc 288:321–333 30. Sprague BL, Trentham-Dietz A, Cronin KA (2012) A sustained decline in postmenopausal hormone use: results from the National Health and Nutrition Examination Survey, 1999–2010. Obstet Gynecol 120:595–603 31. Calle EE, Rodriguez C, Jacobs EJ et al (2002) The American Cancer Society Cancer Prevention Study II Nutrition Cohort. Cancer 94:2490–2501 32. Bergmann MM, Calle EE, Mervis CA et al (1998) Validity of selfreported cancer in a prospective cohort study in comparison with data form state cancer registries. Am J Epidemiol 147:556–562 33. Prat J (2004) Prognostic parameters of endometrial carcinoma. Hum Pathol 35:649–662 34. Cox D (1972) Regression models and life tables. J R Stat Soc B 34:187–202 35. Sherman ME (2000) Theories of endometrial carcinogenesis: a multidisciplinary approach. Mod Path 13:295–308 36. Emons G, Fleckenstein G, Hinney B et al (2000) Hormonal interactions in endometrial cancer. Endocr Relat Cancer 7:227–242 37. Setiawan VW, Yang HP, Pike MC et al (2013) Type I and II endometrial cancers: have they different risk factors? J Clin Oncol 31:2607–2618 38. Cole SR, Platt RW, Schisterman EF et al (2010) Illustrating bias due to conditioning on a collider. Int J Epidemiol 39:417–420 39. Kushi LH, Doyle C, McCullough M et al (2012) American Cancer Society Guidelines on nutrition and physical activity for cancer prevention: reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin 62:30–67
